Hyperspectral sensors collect information from the earth's surface in the form of images with a large number of electromagnetic bands. Accurate classification of hyperspectral images has been one of the hot topics in remote sensing. Spatial information as a complementary source for spectral information helps increase the classification accuracy of hyperspectral images (HSI). Local covariance matrix descriptor (LCMD) is the new spatial-spectral feature generation method for HSI classification. Although the LCMD is easy to use and performs well in HSI classification, it has some limitations, such as discarding the nonlinear relationships between features, which are useful in HSI classification. To address these issues, we propose a local kernel matrix descriptor (LKMD) for the classification of HSIs. In this study, the performance of LCMD is compared with LKMD with two widely used kernels, RBF and polynomial, and final classification results on two real HSIs, Indian Pines and Pavia University, proved the superiority of LKMD over LCMD. تفاصيل المقالة

The classification of Digital Surface Model (DSM) images derived from LiDAR sensors is a challenging task, particularly when distinct ground classes with identical height information must be distinguished. However, DSM images contain valuable spatial information that can be utilized to enhance classification accuracy. This paper proposes a novel strategy, called Multishape Morphological Two-Stage Convolutional Neural Network (MM2CNN), for DSM classification to achieve accurate classified land-cover maps. The proposed method combines the strengths of multishape morphological profiles (MMPs) and a two-stage CNN model as a smart algorithm to effectively discriminate between different land covers from a single-band DSM image. More precisely, the CNN, as a smart method, is used to learn hierarchical rich representations of the data, while the MMPs are used to extract spatial information from the DSM imagery. The approach involves generating MMPs with three structuring elements, training three parallel CNN models, and then stacking and feeding the probability maps to a second-stage CNN to predict the final pixel labels. Experimental results on the Trento benchmark DSM image show that the suggested technique achieves an overall accuracy of 97.32% in a reasonable time, outperforming some other DSM classification methods. The success of the MM2CNN technique demonstrates the potential of integrating MMPs with CNN for precise DSM classification, which has a wide range of applications in environmental investigations. تفاصيل المقالة